DOCSLIB.ORG

Model-Driven Time-Varying Signal Analysis and Its Application to Speech Processing

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Model-Driven Time-Varying Signal Analysis and Its Application to Speech Processing Model-driven Time-varying Signal Analysis and its Application to Speech Processing by Steven Sandoval A Dissertation Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy Approved March 2016 by the Graduate Supervisory Committee: Antonia Papandreou-Suppappola, Chair Julie Liss Pavan Turaga Narayan Kovvali ARIZONA STATE UNIVERSITY May 2016 ABSTRACT This work examines two main areas in model-based time-varying signal process- ing with emphasis in speech processing applications. The first area concentrates on improving speech intelligibility and on increasing the proposed methodologies applica- tion for clinical practice in speech-language pathology. The second area concentrates on signal expansions matched to physical-based models but without requiring inde- pendent basis functions; the significance of this work is demonstrated with speech vowels. A fully automated Vowel Space Area (VSA) computation method is proposed that can be applied to any type of speech. It is shown that the VSA provides an efficient and reliable measure and is correlated to speech intelligibility. A clinical tool that incorporates the automated VSA was proposed for evaluation and treatment to be used by speech language pathologists. Two exploratory studies are performed using two databases by analyzing mean formant trajectories in healthy speech for a wide range of speakers, dialects, and coarticulation contexts. It is shown that phonemes crowded in formant space can often have distinct trajectories, possibly due to accurate perception. A theory for analyzing time-varying signals models with amplitude modulation and frequency modulation is developed. Examples are provided that demonstrate other possible signal model decompositions with independent basis functions and corresponding physical interpretations. The Hilbert transform (HT) and the use of the analytic form of a signal are motivated, and a proof is provided to show that a signal can still preserve desirable mathematical properties without the use of the HT. A visualization of the Hilbert spectrum is proposed to aid in the interpretation. A signal demodulation is proposed and used to develop a modified Empirical Mode Decomposition (EMD) algorithm. i TABLE OF CONTENTS Page LIST OF TABLES . viii LIST OF FIGURES . xi LIST OF ABBREVIATIONS AND ACRONYMS . .xxii CHAPTER 1 INTRODUCTION AND WORK MOTIVATION . 1 1.1 Evaluation of Speech . 1 1.2 Mean Formant Trajectories . 2 1.3 Hilbert Spectral Analysis . 3 1.4 Contributions in the Evaluation of Speech . 5 1.5 Contributions in Hilbert Spectral Analysis . 6 1.6 Report Organization . 7 2 SPEECH QUALITY AND SPEECH INTELLIGIBILITY . 9 2.1 The Evaluation of Speech Quality . 9 2.1.1 Background on Subjective Quality Measures: Mean Opinion Scores . 9 2.1.2 Objective Quality Measures. 10 2.2 The Evaluation of Speech Intelligibility . 14 2.2.1 Standard Assessments of Speech Intelligibility . 15 2.2.2 Automated Assessments of Speech Intelligibility . 16 2.3 Intelligibility Estimation via Automatic Speech Recognition Systems 16 2.3.1 Cepstral Analysis for Speech Characterization . 17 2.3.2 Cepstrum Computation . 18 2.3.3 Mel-Frequency Cepstrum Computation . 19 2.4 Intelligibility Estimation via Feature Mapping . 20 ii CHAPTER Page 2.4.1 Envelope Modulation Spectrum Feature . 21 2.4.2 Long-Term Average Spectrum Feature . 22 2.4.3 Internal ITU-T P.563 Features . 23 2.4.4 Linear Predictive Coding Features. 24 3 AUTOMATIC ASSESSMENT OF VOWEL SPACE AREA. 26 3.1 Motivation of the Proposed Algorithm for Vowel Space Area Auto- matic Assessment Algorithm . 26 3.2 Automatic Assessment Method . 29 3.2.1 Formant Extraction . 29 3.2.2 Filtering . 30 3.2.3 Clustering . 30 3.2.4 Convex Hull / Area Calculation . 30 3.2.5 Stimuli . 31 3.3 Results and Discussion . 31 3.3.1 Performance Analysis . 32 4 SPEECH ASSIST: AN AUGMENTATIVE TOOL FOR PRACTICE IN SPEECH-LANGUAGE PATHOLOGY . 35 4.1 Motivation Proposed for the Mobile Application Suite: Speech Assist 35 4.2 Speech Assist Methods . 35 4.2.1 Automated Vowel Space Area . 35 4.2.2 Automated Diadochokinetic Rate . 36 4.2.3 Automated Perceptual Ratings. 37 4.3 Conceptual Interface . 38 5 MEAN FORMANT TRAJECTORIES . 41 iii CHAPTER Page 5.1 Motivation of Proposed Method for Quantifying Mean Formant Trajectories . 41 5.2 Analysis Study Using the Hillenbrand Database . 44 5.2.1 Analysis Method to Quantify Mean Formant Trajectories . 44 5.2.2 Results and Discussion . 45 5.3 Analysis Study Using TIMIT Database . 48 5.3.1 Analysis Method to Quantity Mean Formant Trajectories . 48 5.3.2 Results and Discussion . 51 6 LATENT SIGNAL ANALYSIS AND THE ANALYTIC SIGNAL . 74 6.1 Motivation for Proposed Latent Signal Analysis . 74 6.2 Background on Instantaneous Frequency . 75 6.3 Latent Signal Analysis . 77 6.4 Hilbert Transform and Analytic Signal . 81 6.4.1 Vakman's Signal Constraints . 82 6.4.2 Analyticity of the Analytic Signal . 83 6.4.3 Gabor's Method . 84 6.5 Relaxing the Constraint of Harmonic Correspondence . 85 6.5.1 Harmonic Correspondence . 85 6.5.2 Analyticity of the Complex Extension . 86 6.5.3 Harmonic Conjugate Functions . 87 6.5.4 Amplitude Modulation{Frequency Modulation Demodulation 88 6.6 Example of a Latent Signal Analysis Problem . 88 6.6.1 Solution Assuming Harmonic Correspondence . 89 6.6.2 Solution Assuming Constant IF . 90 iv CHAPTER Page 6.6.3 Solution Assuming Constant IA . 91 6.7 Discussion. 91 7 HILBERT SPECTRAL ANALYSIS AND THE AM{FM MODEL . 92 7.1 Motivation for Hilbert Spectral Analysis . 92 7.2 The AM{FM Model . 93 7.2.1 Definitions and Assumptions . 93 7.2.2 Monocomponents and Narrowband Components. 96 7.3 Hilbert Spectral Analysis . 98 7.3.1 Two Conventional Ways to Relate a Real Observation to a Latent Signal . 100 7.3.2 Simple Harmonic Component . 101 7.3.3 Superposition of Simple Harmonic Components . 101 7.3.4 The AM Component . 103 7.3.5 Superposition of AM Components . 103 7.3.6 FM Component . 104 7.3.7 Superposition of FM Components . 105 7.3.8 Other AM{FM Models . 105 7.4 Frequency Domain View of Latent Signal Analysis . 107 7.5 Subtleties of the Hilbert Spectrum . 111 7.6 Examples of the Hilbert Spectral Analysis Problem. 113 7.6.1 Periodic Triangle Waveform Example . 114 7.6.2 Sinusoidal FM Example . 115 7.6.3 Remarks . 116 7.7 Visualization of the Hilbert Spectrum . 118 v CHAPTER Page 7.8 Discussion. 120 8 NUMERICAL HILBERT SPECTRAL ANALYSIS ASSUMING INTRIN- SIC MODE FUNCTIONS . 124 8.1 Motivation for Proposed Numerical Methods for Hilbert Spectral Analysis . ..
Load more

Recommended publications
  • Prakriya Documentation Release 0.0.7
    prakriya Documentation Release 0.0.7 Dr. Dhaval Patel Dec 17, 2018 Contents 1 prakriya 3 1.1 Features..................................................3 1.2 Support..................................................3 1.3 Credits..................................................3 2 Installation 5 2.1 Stable release...............................................5 2.2 From sources...............................................5 3 Usage 7 4 Contributing 11 4.1 Types of Contributions.......................................... 11 4.2 Get Started!................................................ 12 4.3 Pull Request Guidelines......................................... 13 4.4 Tips.................................................... 13 5 Credits 15 5.1 Development Lead............................................ 15 5.2 Contributors............................................... 15 6 History 17 6.1 0.0.1 (2017-12-30)............................................ 17 6.2 0.0.2 (2018-01-01)............................................ 17 6.3 0.0.3 (2018-01-02)............................................ 17 6.4 0.0.4 (2018-01-03)............................................ 17 6.5 0.0.5 (2018-01-13)............................................ 17 6.6 0.0.6 (2018-01-16)............................................ 18 6.7 0.0.7 (2018-01-21)............................................ 18 6.8 0.1.0 (2018-12-17)............................................ 18 7 Indices and tables 19 Python Module Index 21 i ii prakriya Documentation, Release 0.0.7 Contents: Contents
    [Show full text]
  • Vedic Accent and Lexicography
    Vedic Accent and Lexicography Felix Rau University of Cologne – Lazarus Project Vedic Accent and Lexicography Lazarus Project: Cologne Sanskrit Lexicon, Project Documentation 2 Felix Rau orcid.org/0000-0003-4167-0601 This work is licensed under the Creative Commons Attribution 4.0 In- ternational License. cite as: Rau, Felix 2017. Vedic Accent and Lexicography. Lazarus Project: Cologne Sanskrit Lexicon, Project Documentation 2. Cologne: Lazarus Project. doi:10.5281/10.5281/zenodo.837826 Lazarus Project (Cologne Sanskrit Lexicon) University of Cologne http://www.cceh.uni-koeln.de/lazarus http://www.sanskrit-lexicon.uni-koeln.de/ 1 Introduction This paper is a preliminary investigation into the problems the representation of the ac- cents of Vedic Sanskrit poses to Sanskrit lexicography. The purpose is to assess the prin- ciples applied in various lexicographic works in the representation of Vedic accents and its relation to the underlying linguistic category as well as traditions of accent marking in different texts. Since the focus is on Sanskrit lexicography, we ignore the complexity of accent marking in manuscripts and the diversity of accent marking across different Indic scripts that were used to write Sanskrit over the ages. We will restrict ourselves to accent marking in Devanagari and Latin script in print, as these two are the relevant systems for virtually all of modern philological Sanskrit lexicography. The complex nature of accent marking in Vedic Sanskrit derives from several facts. Besides the intricacies of the linguistic phenomenon itself (see Kiparsky, 1973, among others), the complexity arises from the fact that different textual or editorial traditions employ structurally different systems for marking Vedic accent.
    [Show full text]
  • Power of Sanskrit
    09/03/2014 WEBPAGE: http://www.translink.profkrishna.com E-mail: [email protected] rofkrishna.com p www. वागतम ् amurthy, Singapore n n Swaagatham N. Krishnamurthy 23 February 2014 www.profkrishna.com Copyright: Dr. N. Krish Consultant, Singapore 1 Acknowledgements and Scope of talk ी गुयो नमः (shri gurub’yo’ namaha) Thanks to Singapore Dakshina Bharatha Brahmana Sabha, and Sri Srinivasan and all members of the rofkrishna.com Sabha Committee for organising this event for me to p p launch my transliteration scheme KrishnaDheva. www. Thanks also to the Sanskrit scholars here, as well as those who have come to learn how to pronounce Sanskrit correctly in English. This talk will not be a religious discourse amurthy, Singapore n This talk will not be a Sanskrit tutoring class This talk will simply be my sharing with you how to: Write down in simple English (KrishnaDheva) any Sanskrit material without special rules, and, Copyright: Dr. N. Krish Read Sanskrit correctly from KrishnaDheva. 2 1 09/03/2014 Starting off The wrong things we say: Sri should be s’ri Siva or Shiva should be s’iva rofkrishna.com p Krishna should be kr+shna www. Visaka should be vis’a’k’a’ Shuklambaradharam should be s’ukla’mbaradh’aram Vrishaba raasi should be vr+shab’a ra’s’ihi amurthy, Singapore n Kowsika gothra should be kaus’ika go’thra … and so on! Copyright: Dr. N. Krish 3 http://sanskritdocuments.org/news/subnews/NASASanskrit.txt Power of Sanskrit – a In ancient India the intention to discover truth was so consuming, that in the process, they discovered perhaps the most perfect tool for fulfilling such a search that the world has ever known – the Sanskrit language.
    [Show full text]
  • Python Module Index 9
    indictransliterationDocumentation Release 0.0.1 sanskrit-programmers Mar 28, 2021 Contents 1 Submodules 3 1.1 indic_transliteration.sanscript......................................3 1.1.1 Submodules...........................................3 1.1.1.1 indic_transliteration.sanscript.schemes........................3 1.1.1.1.1 Submodules.................................3 1.2 indic_transliteration.xsanscript......................................3 1.3 indic_transliteration.detect........................................3 1.3.1 Supported schemes.......................................4 1.4 indic_transliteration.deduplication....................................5 2 Indices and tables 7 Python Module Index 9 Index 11 i ii indictransliterationDocumentation; Release0:0:1 sanscript is the most popular submodule here. Contents 1 indictransliterationDocumentation; Release0:0:1 2 Contents CHAPTER 1 Submodules 1.1 indic_transliteration.sanscript 1.1.1 Submodules 1.1.1.1 indic_transliteration.sanscript.schemes 1.1.1.1.1 Submodules indic_transliteration.sanscript.schemes.roman indic_transliteration.sanscript.schemes.brahmi 1.2 indic_transliteration.xsanscript 1.3 indic_transliteration.detect Example usage: from indic_transliteration import detect detect.detect('pitRRIn') == Scheme.ITRANS detect.detect('pitRRn') == Scheme.HK When handling a Sanskrit string, it’s almost always best to explicitly state its transliteration scheme. This avoids embarrassing errors with words like pitRRIn. But most of the time, it’s possible to infer the encoding from the text itself.
    [Show full text]
  • Vyakarana Documentation Release 0.1
    vyakarana Documentation Release 0.1 Arun Prasad Jul 14, 2017 Contents 1 Background 3 1.1 Introduction...............................................3 1.2 Rule Types................................................4 1.3 Terms and Data..............................................6 1.4 Sounds..................................................8 1.5 asiddha and asiddhavat ......................................... 10 1.6 Glossary................................................. 10 2 Architecture 13 2.1 Design Overview............................................. 13 2.2 Inputs and Outputs............................................ 14 2.3 Modeling Rules............................................. 15 2.4 Selecting Rules.............................................. 17 2.5 Defining Rules.............................................. 17 3 API Reference 19 3.1 API.................................................... 19 Python Module Index 29 i ii vyakarana Documentation, Release 0.1 This is the documentation for Vyakarana, a program that derives Sanskrit words. To get the most out of the documen- tation, you should have a working knowledge of Sanskrit. Important: All data handled by the system is represented in SLP1. SLP1 also uses the following symbols: • '\\' to indicate anudatta¯ • '^' to indicate svarita • '~' to indicate a nasal sound Unmarked vowels are udatta¯ . Contents 1 vyakarana Documentation, Release 0.1 2 Contents CHAPTER 1 Background This is a high-level overview of the Ashtadhyayi and how it works. Introduction This program has two goals: 1. To generate the entire set of forms allowed by the Ashtadhyayi without over- or under-generating. 2. To do so while staying true to the spirit of the Ashtadhyayi. Goal 1 is straightforward, but the “under-generating” is subtle. For some inputs, the Ashtadhyayi can yield multiple results; ideally, we should be able to generate all of them. Goal 2 is more vague. I want to create a program that defines and chooses its rules using the same mechanisms used by the Ashtadhyayi.
    [Show full text]
  • Absence of SUN-Domain Protein Slp1 Blocks Karyogamy and Switches
    Absence of SUN-domain protein Slp1 blocks PNAS PLUS karyogamy and switches meiotic recombination and synapsis from homologs to sister chromatids Christelle Vasniera, Arnaud de Muyta,b, Liangran Zhangc, Sophie Tesséa, Nancy E. Klecknerc,1, Denise Zicklera,1, and Eric Espagnea,1 aInstitut de Génétique et Microbiologie, Unité Mixte de Recherche 8621, Université Paris-Sud, 91405 Orsay, France; bInstitut Curie, 75248 Paris Cedex 05, France; and cDepartment of Molecular and Cellular Biology, Harvard University, Cambridge, MA 02138 Contributed by Nancy E. Kleckner, August 18, 2014 (sent for review June 5, 2014) Karyogamy, the process of nuclear fusion is required for two haploid (7). Blast search of the basidiomycete Cryptococcus neoformans gamete nuclei to form a zygote. Also, in haplobiontic organisms, genome identified four of the known KAR genes; interestingly karyogamy is required to produce the diploid nucleus/cell that then only the kar7 mutant showed a defect in vegetative nuclear enters meiosis. We identify sun like protein 1 (Slp1), member of the movement and hyphae mating but not in premeiotic karyogamy mid–Sad1p, UNC-84–domain ubiquitous family, as essential for kary- (8), likely reflecting the different evolution of ascomycetes and ogamy in the filamentous fungus Sordaria macrospora, thus uncov- basidiomycetes. ering a new function for this protein family. Slp1 is required at the Whereas karyogamy studies of fungi have provided significant last step, nuclear fusion, not for earlier events including nuclear insight into the steps and molecules involved, less is known con- movements, recognition, and juxtaposition. Correspondingly, like cerning the outcome of the remaining nonfused haploid nuclei. other family members, Slp1 localizes to the endoplasmic reticulum Such “twin” meiosis has been analyzed in fission yeast in three and also to its extensions comprising the nuclear envelope.
    [Show full text]
  • A Tool for Transliteration of Bilingual Texts Involving Sanskrit
    A Tool for Transliteration of Bilingual Texts Involving Sanskrit Nikhil Chaturvedi Prof. Rahul Garg IIT Delhi IIT Delhi [email protected] [email protected] Abstract Sanskrit texts are increasingly being written in bilingual and trilingual formats, with Sanskrit paragraphs/shlokas followed by their corresponding English commentary. Sanskrit can also be written in many ways, including multiple encodings like SLP-1 and Velthuis for its romanised form. The need to tackle such code-switching is exacerbated through the requirement to render web pages with multilingual Sanskrit content. These need to automatically detect whether a given text fragment is in Sanskrit, followed by the identification of the form/encoding, further selectively performing transliteration to a user specified script. The Brahmi-derived writing systems of Indian languages are mostly rather similar in structure, but have different letter shapes. These scripts are based on similar phonetic values which allows for easy transliteration. This correspondence forms the basis of the motivation behind deriving a uniform encoding schema that is based on the underlying phonetic value rather than the symbolic representation. The open-source tool developed by us performs this end-to-end detection and transliteration, and achieves an accuracy of 99.1% between SLP-1 and English on a Wikipedia corpus using simple machine learning techniques. 1 Introduction Sanskrit is one of the most ancient languages in India and forms the basis of numerous Indian lan- guages. It is the only known language which has a built-in scheme for pronunciation, word formation and grammar (Maheshwari, 2011). It one of the most used languages of it's time (Huet et al., 2009) and hence encompasses a rich tradition of poetry and drama as well as scientific, technical, philosophical and religious texts.
    [Show full text]
  • SEL-587 Instruction Manual Is Not Available
    Instruction Manual SEL-587-0, -1 Relay Current Differential Relay Overcurrent Relay Instruction Manual 20151105 *PM587-01-NB* © 1995–2015 by Schweitzer Engineering Laboratories, Inc. All rights reserved. All brand or product names appearing in this document are the trademark or registered trademark of their respective holders. No SEL trademarks may be used without written permission. SEL products appearing in this document may be covered by U.S. and Foreign patents. Schweitzer Engineering Laboratories, Inc. reserves all rights and benefits afforded under federal and international copyright and patent laws in its products, including without limitation software, firmware, and documentation. The information in this document is provided for informational use only and is subject to change without notice. Schweitzer Engineering Laboratories, Inc. has approved only the English language document. This product is covered by the standard SEL 10-year warranty. For warranty details, visit www.selinc.com or contact your customer service representative. PM587-01 SEL-587-0, -1 Relay Instruction Manual Date Code 20151105 Table of Contents R.Instruction Manual List of Tables ......................................................................................................................................................vii List of Figures ..................................................................................................................................................... ix Preface..................................................................................................................................................................
    [Show full text]
  • The Cologne Sanskrit Lexicon
    The Cologne Sanskrit Lexicon Felix Rau, Jonathan Blumtritt & Daniel Kölligan University of Cologne Cologne Sanskrit Lexicon 1 11.11.2016 Community & Sustainability for an online collection of 36 dictionaries with hundreds of thousands of entries. Cologne Sanskrit Lexicon 2 11.11.2016 History of the CSL 1994 Dr. Malten initiates the project 1996 Digitization of the Monier-Williams 2003 CSL as a web app 2013 Dr. Malten retires, the DCH becomes responsible, increasing involvement of the community 2013-15 LAZARUS Project Cologne Sanskrit Lexicon 3 11.11.2016 The origins of the CSL • Digitisation of the 1899 edition of Monier-Williams with 31,836 entries • Complex mark-up (later converted to XML) • Originally as a file available for download • Since 2013 as a web app Cologne Sanskrit Lexicon 4 11.11.2016 Cologne Sanskrit Lexicon 5 11.11.2016 Cologne Sanskrit Lexicon 6 11.11.2016 Current CSL • 13 Sanskrit-English dictionaries • 3 English-Sanskrit dictionaries • 2 Sanskrit-French dictionaries • 5 Sanskrit German dictionaries • 1 Sanskrit-Latin dictionary • 2 Sanskrit-Sanskrit dictionaries • 10 specialised (encyclopedic) dictionaries Cologne Sanskrit Lexicon 7 11.11.2016 Current CSL • Internally uses SLP1 to represent Sanskrit • Search input as SLP1, Harvard-Kyoto, and ITRANS • Output in Devanagari, Romanisation, HK, SLP1, and ITRANS • XML Markup (diverse formats, determined by the structure of the printed dictionary) Cologne Sanskrit Lexicon 8 11.11.2016 Cologne Sanskrit Lexicon 9 11.11.2016 Community Cologne Sanskrit Lexicon 10 11.11.2016 1
    [Show full text]
  • Vedic Accent and Lexicography
    ऌ l ̥ and लृ lr ̥ in Sanskrit Lexicography Felix Rau University of Cologne – Lazarus Project ऌ l̥ and लृ lr̥ in Sanskrit Lexicography Lazarus Project: Cologne Sanskrit Lexicon, Project Documentation 1 This work is licensed under the Creative Commons Attribution 4.0 In- ternational License. cite as: Rau, Felix 2017. ऌ l̥ and लृ lr̥ in Sanskrit Lexicography. Lazarus Project: Cologne Sanskrit Lexicon, Project Documentation 1. Cologne: Lazarus Project. doi:10.5281/zenodo.837257 Lazarus Project (Cologne Sanskrit Lexicon) University of Cologne http://www.cceh.uni-koeln.de/lazarus http://www.sanskrit-lexicon.uni-koeln.de/ 1 Introduction This report documents the graphematic representation of the vocalic L (Devanagari: ऌ/ ISO 15919: l̥) and the combination of consonantal L with vocalic R (Devanagari: ल/ृ ISO 15919: lr̥) in Sanskrit lexicography. 2 ऌ l ̥ and लृ lr ̥ Vocalic L is represented by a simple vowel letter ऌ in Devanagari, while the combina- tion of consonantal L with vocalic R is represented by the akṣara ल.ृ Although these two are clearly distinct entities, they were conlated in the original digitisation of the Cologne Sanskrit Lexicon (CSL).1 However, a closer look shows that problems of inconsistent treat- ment of ऌ l̥ and लृ lr̥ have a long and distinguished tradition in the writing systems used in modern Sanskrit lexicography. The relevant writing systems in this context are Devana- gari and Latin script. Of the various Latin-based transliteration systems, four are consid- ered in this report: The ISO standard ISO 15919:2001, the transliteration used in Monier- Williams (1872, 1899), the Sanskrit Library Phonetic Basic encoding scheme (SLP1), and the popular Harvard-Kyoto transliteration system.
    [Show full text]
  • Devanagari Transliteration 1 Devanagari Transliteration
    Devanagari transliteration 1 Devanagari transliteration There are several methods of transliteration from Devanāgarī to the Roman script, which is a process also known as Romanization in the Indian subcontinent. The Hunterian transliteration system is the "national system of romanization in India" and the one officially adopted by the Government of India. IAST is a widely used standard. IAST The International Alphabet of Sanskrit Transliteration (IAST) is a subset of the ISO 15919 standard, used for the transliteration of Sanskrit and Pāḷi into roman script with diacritics. Hunterian transliteration system The Hunterian system was developed in the nineteenth century by William Wilson Hunter, then Surveyor General of India. When it was proposed, it immediately met with opposition from supporters of the earlier practiced non-systematic and often distorting "Sir Roger Dowler method" (an early corruption of Siraj ud-Daulah) of phonetic transcription, which climaxed in a dramatic showdown in an India Council meeting on 28 May 1872 where the new Hunterian method carried the day. The Hunterian method was inherently simpler and extensible to several Indic scripts because it systematized grapheme transliteration, and it came to prevail and gain government and academic acceptance. Opponents of the grapheme transliteration model continued to mount unsuccessful attempts at reversing government policy until the turn of the century, with one critic calling appealing to "the Indian Government to give up the whole attempt at scientific (i.e. Hunterian) transliteration, and decide once and for all in favour of a return to the old phonetic spelling." Over time, the Hunterian method extended in reach to cover several Indic scripts, including Burmese and Tibetan.
    [Show full text]
  • A New Speech Corpus and Modelling Insights Devaraja Adiga1∗ Rishabh Kumar1∗ Amrith Krishna2 [email protected] [email protected] [email protected]
    Automatic Speech Recognition in Sanskrit: A New Speech Corpus and Modelling Insights Devaraja Adiga1∗ Rishabh Kumar1∗ Amrith Krishna2 [email protected] [email protected] [email protected] Preethi Jyothi1 Ganesh Ramakrishnan1 Pawan Goyal3 [email protected] [email protected] [email protected] 1IIT Bombay, Mumbai, India; 2University of Cambridge, UK; 3IIT Kharagpur, WB, India Abstract grapheme-phoneme correspondences. Connected speech leads to phonemic transformations in utter- Automatic speech recognition (ASR) in San- acnes, and in Sanskrit this is faithfully preserved in skrit is interesting, owing to the various lin- guistic peculiarities present in the language. writing as well (Krishna et al., 2018). This is called The Sanskrit language is lexically productive, as Sandhi and is defined as the euphonic assimi- undergoes euphonic assimilation of phones at lation of sounds, i.e., modification and fusion of the word boundaries and exhibits variations in sounds, at or across the boundaries of grammatical spelling conventions and in pronunciations. In units (Matthews, 2007, p. 353). Phonemic orthog- this work, we propose the first large scale study raphy is beneficial for a language, when it comes to of automatic speech recognition (ASR) in San- designing automatic speech recognition Systems skrit, with an emphasis on the impact of unit (ASR), specifically for unit selection at both the selection in Sanskrit ASR. In this work, we release a 78 hour ASR dataset for Sanskrit, Acoustic Model (AM) and Language Model (LM) which faithfully captures several of the linguis- levels. tic characteristics expressed by the language. Regardless of the aforementioned commonali- We investigate the role of different acoustic ties preserved in both the speech and text in San- model and language model units in ASR sys- skrit, designing a large scale ASR system raises tems for Sanskrit.
    [Show full text]